SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2102.01894 · ICCV · 2021

Relaxed Transformer Decoders for Direct Action Proposal Generation

Limin Wang, Jing Tan, Gangshan Wu, Jiaqi Tang

arXiv · PDF · Open in the Atlas

Code that ran

1 of 1 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (1 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 1 has no contract verdict; 0 failures of our instrument; 0 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2102.01894.json · badge: /api/ran/2102.01894.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
MCG-NJU/RTD-Action marked official in Papers with Code
Papers with Code archive
1 ran · 0 no recorded run
FunctionStatusWhere it lives
RTD Ran
ran on a synthesized input · constructed an object; no result was computed
MCG-NJU/RTD-Action/models/rtd.py
code served (permissive licence) · get_code("6e4b96285f344b05")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, RTD: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn
import torch.nn.functional as F

torch.manual_seed(0)

# Minimal position embedding module
class PositionEmbedding(nn.Module):
    def __init__(self, hidden_dim):
        super().__init__()
        self.hidden_dim = hidden_dim
    
    def forward(self, locations):
        # locations shape: (batch_size, T, 1)
        # Return positional embeddings of shape (batch_size, T, hidden_dim)
        batch_size, T, _ = locations.shape
        return torch.randn(batch_size, T, self.hidden_dim)

# Minimal transformer module
class SimpleTransformer(nn.Module):
    def __init__(self, d_model, nhead=2, num_decoder_layers=3):
        super().__init__()
        self.d_model = d_model
        self.nhead = nhead
        
        # Simple transformer decoder for testing
        decoder_layer = nn.TransformerDecoderLayer(
            d_model=d_model,
            nhead=nhead,
            dim_feedforward=256,
            batch_first=True
        )
        self.transformer_decoder = nn.TransformerDecoder(decoder_layer, num_layers=num_decoder_layers)
    
    def forward(self, features, query_embed, pos):
        # features shape: (batch_size, T, hidden_dim)
        # query_embed shape: (num_queries, hidden_dim)
        # pos shape: (batch_size, T, hidden_dim)
        
        batch_size = features.shape[0]
        num_queries = query_embed.shape[0]
        
        # Expand query_embed to batch size
        queries = query_embed.unsqueeze(0).expand(batch_size, -1, -1)
        
        # Run through transformer decoder
        output = self.transformer_decoder(
            tgt=queries,
            memory=features,
            memory_pos=pos
        )
        
        # Return as tuple (output, _) to match expected interface
        # Output shape should be (num_decoder_layers, batch_size, num_queries, d_model)
        # For simplicity, we'll return stacked outputs
        return (output.unsqueeze(0).expand(3, -1, -1, -1),)

# Build arguments
hidden_dim = 256
num_queries = 8
num_classes = 20
batch_size = 2
T = 16

position_embedding = PositionEmbedding(hidden_dim)
transformer = SimpleTransformer(d_model=hidden_dim, nhead=2, num_decoder_layers=3)

# Create sample inputs
locations = torch.randn(batch_size, T, 1)
samples = torch.randn(batch_size, T, 2048)
s_e_scores = torch.rand(batch_size, T, 2)

ARGS = {
    'position_embedding': position_embedding,
    'transformer': transformer,
    'num_classes': num_classes,
    'num_queries': num_queries,
    'stage': 1,
    'aux_loss': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type RTD — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a RTD output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block

Also present in other papers' code

Of 1 samples lifted from this paper's linked repositories, 1 carry a function identity and 0 do not. Of the 1 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 1 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Temporal action proposal generation is an important and challenging task in video understanding, which aims at detecting all temporal segments containing action instances of interest. The existing proposal generation approaches are generally based on pre-defined anchor windows or heuristic bottom-up boundary matching strategies. This paper presents a simple and efficient framework (RTD-Net) for direct action proposal generation, by re-purposing a Transformer-alike architecture. To tackle the essential visual difference between time and space, we make three important improvements over the original transformer detection framework (DETR). First, to deal with slowness prior in videos, we replace the original Transformer encoder with a boundary attentive module to better capture long-range temporal information. Second, due to the ambiguous temporal boundary and relatively sparse annotations, we present a relaxed matching scheme to relieve the strict criteria of single assignment to each groundtruth. Finally, we devise a three-branch head to further improve the proposal confidence estimation by explicitly predicting its completeness. Extensive experiments on THUMOS14 and ActivityNet-1.3 benchmarks demonstrate the effectiveness of RTD-Net, on both tasks of temporal action proposal generation and temporal action detection. Moreover, due to its simplicity in design, our framework is more efficient than previous proposal generation methods, without non-maximum suppression post-processing. The code and models are made available at https://github.com/ MCG-NJU/RTD-Action.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2102.01894")
get_code_for_paper("2102.01894")
have("2102.01894")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2102.01894.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2102.01894.svg)](https://syntology.ai/paper/2102.01894)

Connect an agent — have() is free.